Tensorflow on ARC - Revision history

Dmitri: /* Installing with Conda */

2024-03-07T22:44:26Z

Installing with Conda

@@ Line 29: / Line 29: @@
 == Installing with Conda ==
-We will install '''Tensorflow-GPU''' into its own '''conda environment'''.
+We will install '''Tensorflow''' into its own '''conda environment'''.
-It is '''very important''' to create the environment with '''python''' and '''tensorflow-gpu''' in the same command.
+Now we activate our new environment
-This way '''conda''' can select the best '''tensorflow-gpu''' and '''python''' combination.
+  $ conda activate tensorflow
-Once it is done,
+Check for the available versions of <code>pytorch</code> on the <code>conda-forge</code> channel and decide on what you want to install.
-activate your '''tensorflow''' ennvironment:
+Please make sure that there is '''GPU''' or '''cuda''' build for it (3rd column in the output table).
- $ conda activate tensorflow
+<pre>
 == Installing with PIP ==

2024-01-08T23:33:34Z

Installing with PIP

← Older revision		Revision as of 23:33, 8 January 2024
Line 52:		Line 52:
	....		....
	</pre>		</pre>
	If the installation has been successful, you now have the required versions of Python and Pip installed. \		If the installation has been successful, you now have the required versions of Python and Pip installed.
	You can now follow the instructions from the manual.		You can now follow the instructions from the manual.
	<pre>		<pre>

2024-01-08T23:30:44Z

Installing with PIP

← Older revision		Revision as of 23:30, 8 January 2024
Line 58:		Line 58:
	....		....
	</pre>		</pre>
	After this installation is finished ~~the~~ Tensorflow is ready for testing.		After this installation is finished Tensorflow is ready for testing.

	== Testing ==		== Testing ==

2024-01-08T23:29:12Z

Installing Tensorflow

← Older revision		Revision as of 23:29, 8 January 2024
Line 21:		Line 21:

	Once you have your own '''Conda''', activate it with		Once you have your own '''Conda''', activate it with
	$ ~/software/init-conda		$ source ~/software/init-conda

	Here you have two choices how to install Tensorflow		Here you have two choices how to install Tensorflow

2024-01-08T23:28:55Z

Installing Tensorflow

← Older revision		Revision as of 23:28, 8 January 2024
Line 18:		Line 18:
	You will need a working local '''Conda''' install in your home directory first.		You will need a working local '''Conda''' install in your home directory first.
	If you do not have it yet, plaese follow [[Conda_on_ARC#Installing_Conda \| these instructions]]		If you do not have it yet, plaese follow [[Conda_on_ARC#Installing_Conda \| these instructions]]
	to have it ~~isntalled~~.		to have it installed.

	Once you have your own '''Conda''', activate it with		Once you have your own '''Conda''', activate it with

2024-01-08T23:26:16Z

Testing

← Older revision		Revision as of 23:26, 8 January 2024
Line 74:		Line 74:
	To deactivate the environment use the		To deactivate the environment use the
	(tensorflow) $ conda deactivate		(tensorflow) $ conda deactivate
	~~command.~~

	And leave Conda		command. And leave Conda
	(base) $ conda deactivate		(base) $ conda deactivate

2024-01-08T23:25:55Z

Testing

← Older revision		Revision as of 23:25, 8 January 2024
Line 64:		Line 64:
	You can test with the <code>tensorflow-test.py</code> script shown below.		You can test with the <code>tensorflow-test.py</code> script shown below.
	Copy and paste the text into a file and run if from the command line:		Copy and paste the text into a file and run if from the command line:
	$ python tensorflow-test.py		(tensorflow) $ python tensorflow-test.py

	If you try this on the login node, it should tell you that GPUs are not available.		If you try this on the login node, it should tell you that GPUs are not available.
Line 73:		Line 73:

	To deactivate the environment use the		To deactivate the environment use the
	$ conda deactivate		(tensorflow) $ conda deactivate
	command.		command.

			And leave Conda
			(base) $ conda deactivate

	=== Test script ===		=== Test script ===

2024-01-08T23:24:37Z

Installing Tensorflow

← Older revision		Revision as of 23:24, 8 January 2024
Line 23:		Line 23:
	$ ~/software/init-conda		$ ~/software/init-conda

	Here you have ~~to decide if you want~~ to install Tensorflow (1) using Conda or (2) how it is explained on the Tensorflow site, using <code>pip</code>.		Here you have two choices how to install Tensorflow
			(1) using Conda or
			(2) how it is explained on the Tensorflow site, using <code>pip</code>.

	== Installing with Conda ==		== Installing with Conda ==

2024-01-08T23:22:55Z

Installing Tensorflow

@@ Line 19: / Line 19: @@
 If you do not have it yet, plaese follow [[Conda_on_ARC#Installing_Conda | these instructions]]
 to have it isntalled.
 Once you have your own '''Conda''', activate it with
   $ ~/software/init-conda
 We will install '''Tensorflow-GPU''' into its own '''conda environment'''.
@@ Line 33: / Line 36: @@
 activate your '''tensorflow''' ennvironment:
   $ conda activate tensorflow
 You can test with the <code>tensorflow-test.py</code> script shown below.

2024-01-08T23:10:59Z

Test script

← Older revision		Revision as of 23:10, 8 January 2024
Line 59:		Line 59:
	os.environ['TF_XLA_FLAGS'] = '--tf_xla_enable_xla_devices'		os.environ['TF_XLA_FLAGS'] = '--tf_xla_enable_xla_devices'
	os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'		os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
	~~# ------------------------------------------~~
	sep = "=" * 80		sep = "=" * 80
	# ------------------------------------------		# ------------------------------------------